Re: [PATCH RFC net-next 1/2] net: Reference bpf_redirect_info via task_struct on PREEMPT_RT.

[Sebastian Andrzej Siewior <bigeasy@linutronix.de>]

On 2024-02-22 11:10:44 [+0100], Jesper Dangaard Brouer wrote:
> > Ethtool(eth1    ) stat:     14158562 (     14,158,562) <= tx_packets /sec
> > Ethtool(eth1    ) stat:     14158685 (     14,158,685) <= tx_pkts_nic /sec
> > 
> > looks like a small improvement… It is not your 15 but close. -t2 does
> > improve the situation.
> 
> You cannot reach 15Mpps on 10Gbit/s as wirespeed for 10G is 14.88Mpps.

Oh, my bad.

> Congratulations, I think this 14.15 Mpps is as close to wirespeed as it
> possible on your hardware.
> 
> BTW what CPU are you using?

"Intel(R) Xeon(R) CPU E7-8890 v3 @ 2.50GHz"
The "performance" governor is used, I lowered the number of CPUs and
disabled SMT.

> > There is a warning from DMA mapping code but ;)
> 
> It is a warning from IOMMU code?
> It usually means there is a real DMA unmap bug (which we should fix).

Not sure, I don't think so:
| ------------[ cut here ]------------
| ehci-pci 0000:00:1a.0: DMA addr 0x0000000105016ce8+8 overflow (mask ffffffff, bus limit 0).
| WARNING: CPU: 0 PID: 1029 at kernel/dma/direct.h:105 dma_map_page_attrs+0x1e8/0x1f0
| RIP: 0010:dma_map_page_attrs+0x1e8/0x1f0
| Call Trace:
|  <TASK>
|  usb_hcd_map_urb_for_dma+0x1b0/0x4d0
|  usb_hcd_submit_urb+0x342/0x9b0
|  usb_start_wait_urb+0x50/0xc0
|  usb_control_msg+0xc8/0x110
|  get_bMaxPacketSize0+0x5a/0xb0

and USB isn't working. I *think* it is the "memory above 4G" thing, not
sure where it took the wrong turn.

> --Jesper

Sebastian